Tag Collection
2 published prompts tagged “智能体安全”. Browse by scenario and copy in one click.
Few indexable prompts for this tag; the page may be noindex.
2 prompts
This prompt guides a comprehensive safety and control review of an agent system across dimensions of human control, goal understanding, security, transparency, and privacy, requiring a structured evaluation report.
Design and execute adversarial test campaigns against AI agent systems—including single/multi-agent, MCP servers, skill ecosystems, and long-horizon autonomous workflows. Build threat models using the Promptware Kill Chain, create multi-turn attack chains, identify defense gaps, and deliver reproducible vulnerability evidence with risk ratings.
They are reusable LLM prompt templates labeled with “智能体安全” in Easy Prompt, selected for practical workflows and clear structure.
Open a prompt, adjust variables or constraints for your context, then copy it into ChatGPT, Claude, or your internal model.
This page lists published prompts with the tag. Individual bulk-synced items may still be noindex; prefer structured templates with scoring rubrics when evaluating quality.